please note that I am not a developer, but am vetting the research methodology of an experiment using Firebase A/B testing (using a contractor for app dev)- I need to better understand how randomization into groups works (beyond what's described on Firebase's remote config and A/B testing documentation).
100% of users will be targeted for the experiment, and I have the configuration between treatment and control set to 50/50, but would like to know if it’s:
- List item
“simple” randomization (that is, when a new user joins, they are selected independent of existing group assignment ratios, e.g., via a random number generator between 0-1 where if a user’s random number >0.5 they’re in treatment, and <0.5 they’re in control) OR
- List item
“balanced” randomization, where Firebase sorts users into groups to achieve a 50/50 split; that is, group assignment is not independent of the existing group assignment ratio (e.g., if there are currently 100 users in treatment and 98 users in control, it will assign the new user to control to get closer to 50/50)
Thank you so much in advance! Btw, if you do have specific lines of code I can confirm back with my dev contractor to confirm randomization protocol, that would be helpful.
